Merge PDF

1. Choose the PDFs on your device. Drop the files onto the panel above, or press Choose PDFs. Up to 20 at a time, 20 MB each, and they stay where they are — there is no transfer step.
2. Let each file report its pages. Every file is opened as it lands and its page count is shown on its row. A file that is locked, damaged or empty is named right there instead of failing later.
3. Put the documents in order. The number beside each row is the position it holds in the finished file. Use the Move up and Move down buttons on a row to shift it, and the remove button to drop one out.
4. Pick pages, if you only want some. Leave a row on All pages to take the whole file. Switch it to Some pages and type numbers or ranges — 1-3, 7 — to take only those, in the order you typed them.
5. Press Combine and download. Your device copies the pages into one new document and prints what it holds and what it weighs. Download it, open it to check the order, and run it again if you want it different.

Combining PDFs is normally done by taking a copy of your files first and doing the work on a machine somewhere else. That is a reasonable way to build software and a poor fit for what people actually merge — signed contracts, bank statements, payslips, medical letters and scans of a passport. Those are the files where the question is not how fast the tool is, but who else ends up holding a copy.

Here the code that reads and writes a PDF arrives with the page and runs in this browser tab. Your device opens the files, copies the pages across and hands you the finished document, so there is no copy of your paperwork anywhere but on your own computer. The only thing that ever travels is the code, in one direction, on its way to you.

A PDF page is not a picture. It is a small program — draw this text in this font at this position, paint this image into this rectangle — stored as a compressed stream alongside the resources it refers to. Combining files moves that whole object across into a new document exactly as it is.

So nothing is decoded and nothing is re-encoded. Text stays selectable text rather than becoming a picture of text. A scan embedded in a page arrives as the same image, byte for byte, at the same resolution. The page size and any rotation the page carries come across with it, which is what keeps a landscape sheet landscape and a page a scanner turned 90 degrees still turned. The finished file weighs about what its parts weighed added together, because that is essentially what it is.

A password-protected PDF has encrypted page contents. The library used here throws an error on one, and its error message suggests a flag that carries on regardless. That flag was tried and measured: the file opens, the pages copy, the merge reports success — and the pages that land are still encrypted, so a reader shows them blank.

A tool that did that would hand somebody an empty contract and tell them it worked. So the flag is never used, there is a test that re-proves why, and a locked file is refused by name with the only fix there is: take the password off in the app that made the file, save a copy, and add that copy instead.

Each row can take the whole file or a list of pages. The box accepts single numbers and ranges together — 1-3, 7 — and the numbers are the ones your reader shows, counting from 1.

The order you type is the order the pages land in, which makes two useful things fall out of the same rule rather than needing switches of their own. A descending range reverses a section, so 5-3 gives you pages 5, 4 and 3 in that order — the answer to a stack that went through a scanner the wrong way up. And naming a page twice puts it in twice, so a cover page can also close the document. Anything outside the file is caught as you type, against the page count that file actually reported.

• A visa or mortgage application that wants every document as a single file.
• Six months of bank statements that arrived as one PDF per month.
• A contract, its annexes and a signature page that have to go back together.
• An invoice and its receipts, submitted as one attachment to an expense system.
• Scans made in two sittings, where the second half needs to follow the first.
• A report assembled from sections different people wrote separately.

Up to 20 PDFs at once, 20 MB per file and 80 MB across the set. PDFs only, and each one is checked by its first bytes rather than by its name — so a file renamed to end in .pdf is refused for what it really is.

Past those numbers the binding limit is the device in your hand, not a policy. It is a generous limit here, though, because a merge costs far less than an image job: no page is ever turned into pixels, so the whole thing is priced in bytes, and the panel refuses in plain words before it starts if this phone or laptop genuinely cannot hold it. It is the honest answer — there is nowhere else for the work to go.